* common.mk (COMMONOBJS): transcode_data_*.c moved under enc/trans.

* transcode_data.h (rb_transcoding, rb_transcoder): prefixed.

* transcode.c (rb_register_transcoder, rb_declare_transcoder): split
  declaration and registration.  [ruby-dev:32704]

* transcode.c (transcode_dispatch): autoload pre-declared transcoder.

* transcode.c (str_transcode): use rb_define_dummy_encoding().

* transcode.c (Init_transcode): initialize transcoder tables.

* enc/trans/single_byte.c, enc/trans/japanese.c: moved from top.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@14666 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

6 files changed, 425 insertions, 241 deletions

diff --git a/ChangeLog b/ChangeLog
index ceaee71160..b8b15e1b06 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,20 @@
+Tue Dec 25 14:57:00 2007  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* common.mk (COMMONOBJS): transcode_data_*.c moved under enc/trans.
+
+	* transcode_data.h (rb_transcoding, rb_transcoder): prefixed.
+
+	* transcode.c (rb_register_transcoder, rb_declare_transcoder): split
+	  declaration and registration.  [ruby-dev:32704]
+
+	* transcode.c (transcode_dispatch): autoload pre-declared transcoder.
+
+	* transcode.c (str_transcode): use rb_define_dummy_encoding().
+
+	* transcode.c (Init_transcode): initialize transcoder tables.
+
+	* enc/trans/single_byte.c, enc/trans/japanese.c: moved from top.
+
 Tue Dec 25 14:20:13 2007  Yukihiro Matsumoto  <matz@ruby-lang.org>
 
 	* lib/mkmf.rb (map_dir): should generate path including $top_srcdir.
diff --git a/common.mk b/common.mk
index 634fdcaf54..fd5c366b5c 100644
--- a/common.mk
+++ b/common.mk
@@ -60,8 +60,6 @@ COMMONOBJS    = array.$(OBJEXT) \
 		struct.$(OBJEXT) \
 		time.$(OBJEXT) \
 		transcode.$(OBJEXT) \
-		transcode_data_one_byte.$(OBJEXT) \
-		transcode_data_japanese.$(OBJEXT) \
 		util.$(OBJEXT) \
 		variable.$(OBJEXT) \
 		version.$(OBJEXT) \
@@ -548,8 +546,6 @@ thread.$(OBJEXT): {$(VPATH)}thread.c {$(VPATH)}eval_intern.h \
   {$(VPATH)}signal.h {$(VPATH)}st.h {$(VPATH)}dln.h
 transcode.$(OBJEXT): {$(VPATH)}transcode.c {$(VPATH)}transcode_data.h {$(VPATH)}ruby.h {$(VPATH)}config.h \
   {$(VPATH)}defines.h {$(VPATH)}intern.h {$(VPATH)}missing.h  {$(VPATH)}encoding.h
-transcode_data_one_byte.$(OBJEXT): {$(VPATH)}transcode_data_one_byte.c {$(VPATH)}transcode_data.h
-transcode_data_japanese.$(OBJEXT): {$(VPATH)}transcode_data_japanese.c {$(VPATH)}transcode_data.h
 cont.$(OBJEXT):  {$(VPATH)}cont.c {$(VPATH)}eval_intern.h \
   {$(VPATH)}ruby.h {$(VPATH)}vm_core.h {$(VPATH)}id.h {$(VPATH)}config.h \
   {$(VPATH)}defines.h {$(VPATH)}intern.h {$(VPATH)}missing.h \
diff --git a/transcode_data_japanese.c b/enc/trans/japanese.c
index 860b824221..3e03f2c969 100644
--- a/transcode_data_japanese.c
+++ b/enc/trans/japanese.c
@@ -1,4 +1,3 @@
-#define TRANSCODE_DATA
 #include "transcode_data.h"
 
 static const unsigned char
@@ -4418,11 +4417,16 @@ from_SHIFT_JIS_infos[108] = {
      &from_SHIFT_JIS_FA, &from_SHIFT_JIS_FB,
      &from_SHIFT_JIS_FC,              UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_SHIFT_JIS = {
+static const BYTE_LOOKUP
+from_SHIFT_JIS = {
     from_SHIFT_JIS_offsets,
     from_SHIFT_JIS_infos
 };
+static rb_transcoder
+rb_from_SHIFT_JIS = {
+    "UTF-8", "SHIFT_JIS", &from_SHIFT_JIS, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_SHIFT_JIS_C2_offsets[64] = {
@@ -13343,11 +13347,16 @@ to_SHIFT_JIS_infos[17] = {
      &to_SHIFT_JIS_E7, &to_SHIFT_JIS_E8, &to_SHIFT_JIS_E9, &to_SHIFT_JIS_EF,
                 UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_SHIFT_JIS = {
+static const BYTE_LOOKUP
+to_SHIFT_JIS = {
     to_SHIFT_JIS_offsets,
     to_SHIFT_JIS_infos
 };
+static rb_transcoder
+rb_to_SHIFT_JIS = {
+    "SHIFT_JIS", "UTF-8", &to_SHIFT_JIS, 2, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_EUC_JP_8E_offsets[256] = {
@@ -18112,11 +18121,16 @@ from_EUC_JP_infos[85] = {
      &from_EUC_JP_F9, &from_EUC_JP_FA, &from_EUC_JP_FB, &from_EUC_JP_FC,
                UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_EUC_JP = {
+static const BYTE_LOOKUP
+from_EUC_JP = {
     from_EUC_JP_offsets,
     from_EUC_JP_infos
 };
+static rb_transcoder
+rb_from_EUC_JP = {
+    "UTF-8", "EUC-JP", &from_EUC_JP, 3, 0,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 to_EUC_JP_C2_infos[14] = {
@@ -23612,11 +23626,16 @@ to_EUC_JP_infos[17] = {
      &to_EUC_JP_E7, &to_EUC_JP_E8, &to_EUC_JP_E9, &to_EUC_JP_EF,
              UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_EUC_JP = {
+static const BYTE_LOOKUP
+to_EUC_JP = {
     to_SHIFT_JIS_offsets,
     to_EUC_JP_infos
 };
+static rb_transcoder
+rb_to_EUC_JP = {
+    "EUC_JP", "UTF-8", &to_EUC_JP, 2, 1,
+    NULL, NULL,
+};
 
 #define ISO_2022_ENCODING(escseq, byte) ((escseq<<8)|byte)
 enum ISO_2022_ESCSEQ {
@@ -23699,11 +23718,11 @@ get_iso_2022_mode(char **in_pos)
     return new_mode;
 }
 
-void
+static void
 from_iso_2022_jp_transcoder_preprocessor(char **in_pos, char **out_pos,
-				     char *in_stop, char *out_stop,
-				     transcoder *my_transcoder,
-				     transcoding *my_transcoding)
+					 char *in_stop, char *out_stop,
+					 rb_transcoder *my_transcoder,
+					 rb_transcoding *my_transcoding)
 {
     char *in_p = *in_pos, *out_p = *out_pos;
     int cur_mode = ISO_2022_GZ_ASCII;
@@ -23779,11 +23798,11 @@ select_iso_2022_mode(char **out_pos, int new_mode)
     return new_mode;
 }
 
-void
+static void
 to_iso_2022_jp_transcoder_postprocessor(char **in_pos, char **out_pos,
-				     char *in_stop, char *out_stop,
-				     transcoder *my_transcoder,
-				     transcoding *my_transcoding)
+					char *in_stop, char *out_stop,
+					rb_transcoder *my_transcoder,
+					rb_transcoding *my_transcoding)
 {
     char *in_p = *in_pos, *out_p = *out_pos;
     int cur_mode = ISO_2022_GZ_ASCII, new_mode = 0;
@@ -23824,3 +23843,26 @@ to_iso_2022_jp_transcoder_postprocessor(char **in_pos, char **out_pos,
     *in_pos  = in_p;
     *out_pos = out_p;
 }
+
+static rb_transcoder
+rb_from_ISO_2022_JP = {
+    "ISO-2022-JP", "UTF-8", &from_EUC_JP, 8, 0,
+    &from_iso_2022_jp_transcoder_preprocessor, NULL,
+};
+
+static rb_transcoder
+rb_to_ISO_2022_JP = {
+    "UTF-8", "ISO-2022-JP", &to_EUC_JP, 8, 1,
+    NULL, &to_iso_2022_jp_transcoder_postprocessor,
+};
+
+void
+Init_japanese(void)
+{
+    rb_register_transcoder(&rb_from_SHIFT_JIS);
+    rb_register_transcoder(&rb_from_EUC_JP);
+    rb_register_transcoder(&rb_to_SHIFT_JIS);
+    rb_register_transcoder(&rb_to_EUC_JP);
+    rb_register_transcoder(&rb_from_ISO_2022_JP);
+    rb_register_transcoder(&rb_to_ISO_2022_JP);
+}
diff --git a/transcode_data_one_byte.c b/enc/trans/single_byte.c
index f050bfeb4f..ca4006d542 100644
--- a/transcode_data_one_byte.c
+++ b/enc/trans/single_byte.c
@@ -1,4 +1,3 @@
-#define TRANSCODE_DATA
 #include "transcode_data.h"
 
 static const unsigned char
@@ -65,11 +64,16 @@ from_ISO_8859_1_infos[129] = {
      o2(0xC3,0xBB), o2(0xC3,0xBC), o2(0xC3,0xBD), o2(0xC3,0xBE),
      o2(0xC3,0xBF),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_1 = {
+static const BYTE_LOOKUP
+from_ISO_8859_1 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_1_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_1 = {
+    "ISO-8859-1", "UTF-8", &from_ISO_8859_1, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_1_C2_offsets[64] = {
@@ -158,11 +162,16 @@ to_ISO_8859_1_infos[4] = {
                  NOMAP, &to_ISO_8859_1_C2,
      &to_ISO_8859_1_C3,             UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_1 = {
+static const BYTE_LOOKUP
+to_ISO_8859_1 = {
     to_ISO_8859_1_offsets,
     to_ISO_8859_1_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_1 = {
+    "UTF-8", "ISO-8859-1", &to_ISO_8859_1, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_2_infos[129] = {
@@ -200,11 +209,16 @@ from_ISO_8859_2_infos[129] = {
      o2(0xC5,0xB1), o2(0xC3,0xBC), o2(0xC3,0xBD), o2(0xC5,0xA3),
      o2(0xCB,0x99),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_2 = {
+static const BYTE_LOOKUP
+from_ISO_8859_2 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_2_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_2 = {
+    "ISO-8859-2", "UTF-8", &from_ISO_8859_2, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_2_C2_offsets[64] = {
@@ -351,11 +365,16 @@ to_ISO_8859_2_infos[7] = {
      &to_ISO_8859_2_C5, &to_ISO_8859_2_CB,
                  UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_2 = {
+static const BYTE_LOOKUP
+to_ISO_8859_2 = {
     to_ISO_8859_2_offsets,
     to_ISO_8859_2_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_2 = {
+    "UTF-8", "ISO-8859-2", &to_ISO_8859_2, 1, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_ISO_8859_3_offsets[256] = {
@@ -410,11 +429,16 @@ from_ISO_8859_3_infos[123] = {
      o2(0xC3,0xBA), o2(0xC3,0xBB), o2(0xC3,0xBC), o2(0xC5,0xAD),
      o2(0xC5,0x9D), o2(0xCB,0x99),         UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_3 = {
+static const BYTE_LOOKUP
+from_ISO_8859_3 = {
     from_ISO_8859_3_offsets,
     from_ISO_8859_3_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_3 = {
+    "ISO-8859-3", "UTF-8", &from_ISO_8859_3, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_3_C2_offsets[64] = {
@@ -536,11 +560,16 @@ to_ISO_8859_3_infos[7] = {
      &to_ISO_8859_3_C5, &to_ISO_8859_3_CB,
                  UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_3 = {
+static const BYTE_LOOKUP
+to_ISO_8859_3 = {
     to_ISO_8859_2_offsets,
     to_ISO_8859_3_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_3 = {
+    "UTF-8", "ISO-8859-3", &to_ISO_8859_3, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_4_infos[129] = {
@@ -578,11 +607,16 @@ from_ISO_8859_4_infos[129] = {
      o2(0xC3,0xBB), o2(0xC3,0xBC), o2(0xC5,0xA9), o2(0xC5,0xAB),
      o2(0xCB,0x99),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_4 = {
+static const BYTE_LOOKUP
+from_ISO_8859_4 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_4_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_4 = {
+    "ISO-8859-4", "UTF-8", &from_ISO_8859_4, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_4_C2_offsets[64] = {
@@ -708,11 +742,16 @@ to_ISO_8859_4_infos[7] = {
      &to_ISO_8859_4_C5, &to_ISO_8859_4_CB,
                  UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_4 = {
+static const BYTE_LOOKUP
+to_ISO_8859_4 = {
     to_ISO_8859_2_offsets,
     to_ISO_8859_4_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_4 = {
+    "UTF-8", "ISO-8859-4", &to_ISO_8859_4, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_5_infos[129] = {
@@ -782,11 +821,16 @@ from_ISO_8859_5_infos[129] = {
           o2(0xC2,0xA7),      o2(0xD1,0x9E),
           o2(0xD1,0x9F),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_5 = {
+static const BYTE_LOOKUP
+from_ISO_8859_5 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_5_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_5 = {
+    "ISO-8859-5", "UTF-8", &from_ISO_8859_5, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_5_C2_offsets[64] = {
@@ -928,11 +972,16 @@ to_ISO_8859_5_infos[6] = {
      &to_ISO_8859_5_D0, &to_ISO_8859_5_D1,
      &to_ISO_8859_5_E2,             UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_5 = {
+static const BYTE_LOOKUP
+to_ISO_8859_5 = {
     to_ISO_8859_5_offsets,
     to_ISO_8859_5_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_5 = {
+    "UTF-8", "ISO-8859-5", &to_ISO_8859_5, 1, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_ISO_8859_6_offsets[256] = {
@@ -978,11 +1027,16 @@ from_ISO_8859_6_infos[85] = {
      o2(0xD9,0x8F), o2(0xD9,0x90), o2(0xD9,0x91), o2(0xD9,0x92),
              UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_6 = {
+static const BYTE_LOOKUP
+from_ISO_8859_6 = {
     from_ISO_8859_6_offsets,
     from_ISO_8859_6_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_6 = {
+    "ISO-8859-6", "UTF-8", &from_ISO_8859_6, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_6_C2_offsets[64] = {
@@ -1079,11 +1133,16 @@ to_ISO_8859_6_infos[5] = {
      &to_ISO_8859_6_D8, &to_ISO_8859_6_D9,
                  UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_6 = {
+static const BYTE_LOOKUP
+to_ISO_8859_6 = {
     to_ISO_8859_6_offsets,
     to_ISO_8859_6_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_6 = {
+    "UTF-8", "ISO-8859-6", &to_ISO_8859_6, 1, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_ISO_8859_7_offsets[256] = {
@@ -1171,11 +1230,16 @@ from_ISO_8859_7_infos[127] = {
           o2(0xCF,0x8D),      o2(0xCF,0x8E),
                   UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_7 = {
+static const BYTE_LOOKUP
+from_ISO_8859_7 = {
     from_ISO_8859_7_offsets,
     from_ISO_8859_7_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_7 = {
+    "ISO-8859-7", "UTF-8", &from_ISO_8859_7, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_7_C2_offsets[64] = {
@@ -1352,11 +1416,16 @@ to_ISO_8859_7_infos[7] = {
      &to_ISO_8859_7_CF, &to_ISO_8859_7_E2,
                  UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_7 = {
+static const BYTE_LOOKUP
+to_ISO_8859_7 = {
     to_ISO_8859_7_offsets,
     to_ISO_8859_7_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_7 = {
+    "UTF-8", "ISO-8859-7", &to_ISO_8859_7, 1, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_ISO_8859_8_offsets[256] = {
@@ -1427,11 +1496,16 @@ from_ISO_8859_8_infos[94] = {
           o2(0xD7,0xAA), o3(0xE2,0x80,0x8E),
      o3(0xE2,0x80,0x8F),              UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_8 = {
+static const BYTE_LOOKUP
+from_ISO_8859_8 = {
     from_ISO_8859_8_offsets,
     from_ISO_8859_8_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_8 = {
+    "ISO-8859-8", "UTF-8", &from_ISO_8859_8, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_8_C2_offsets[64] = {
@@ -1567,11 +1641,16 @@ to_ISO_8859_8_infos[6] = {
      &to_ISO_8859_8_C3, &to_ISO_8859_8_D7,
      &to_ISO_8859_8_E2,             UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_8 = {
+static const BYTE_LOOKUP
+to_ISO_8859_8 = {
     to_ISO_8859_8_offsets,
     to_ISO_8859_8_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_8 = {
+    "UTF-8", "ISO-8859-8", &to_ISO_8859_8, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_9_infos[129] = {
@@ -1609,11 +1688,16 @@ from_ISO_8859_9_infos[129] = {
      o2(0xC3,0xBB), o2(0xC3,0xBC), o2(0xC4,0xB1), o2(0xC5,0x9F),
      o2(0xC3,0xBF),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_9 = {
+static const BYTE_LOOKUP
+from_ISO_8859_9 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_9_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_9 = {
+    "ISO-8859-9", "UTF-8", &from_ISO_8859_9, 2, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_9_C3_offsets[64] = {
@@ -1706,11 +1790,16 @@ to_ISO_8859_9_infos[6] = {
      &to_ISO_8859_9_C3, &to_ISO_8859_9_C4,
      &to_ISO_8859_9_C5,             UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_9 = {
+static const BYTE_LOOKUP
+to_ISO_8859_9 = {
     to_ISO_8859_9_offsets,
     to_ISO_8859_9_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_9 = {
+    "UTF-8", "ISO-8859-9", &to_ISO_8859_9, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_10_infos[129] = {
@@ -1780,11 +1869,16 @@ from_ISO_8859_10_infos[129] = {
           o2(0xC3,0xBD),      o2(0xC3,0xBE),
           o2(0xC4,0xB8),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_10 = {
+static const BYTE_LOOKUP
+from_ISO_8859_10 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_10_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_10 = {
+    "ISO-8859-10", "UTF-8", &from_ISO_8859_10, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_10_C2_offsets[64] = {
@@ -1932,11 +2026,16 @@ to_ISO_8859_10_infos[7] = {
      &to_ISO_8859_10_C5, &to_ISO_8859_10_E2,
                   UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_10 = {
+static const BYTE_LOOKUP
+to_ISO_8859_10 = {
     to_ISO_8859_10_offsets,
     to_ISO_8859_10_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_10 = {
+    "UTF-8", "ISO-8859-10", &to_ISO_8859_10, 1, 1,
+    NULL, NULL,
+};
 
 static const unsigned char
 from_ISO_8859_11_offsets[256] = {
@@ -2021,11 +2120,16 @@ from_ISO_8859_11_infos[122] = {
      o3(0xE0,0xB9,0x99), o3(0xE0,0xB9,0x9A),
      o3(0xE0,0xB9,0x9B),              UNDEF,
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_11 = {
+static const BYTE_LOOKUP
+from_ISO_8859_11 = {
     from_ISO_8859_11_offsets,
     from_ISO_8859_11_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_11 = {
+    "ISO-8859-11", "UTF-8", &from_ISO_8859_11, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_11_C2_offsets[64] = {
@@ -2149,11 +2253,16 @@ to_ISO_8859_11_infos[4] = {
                   NOMAP, &to_ISO_8859_11_C2,
      &to_ISO_8859_11_E0,              UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_11 = {
+static const BYTE_LOOKUP
+to_ISO_8859_11 = {
     to_ISO_8859_11_offsets,
     to_ISO_8859_11_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_11 = {
+    "UTF-8", "ISO-8859-11", &to_ISO_8859_11, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_13_infos[129] = {
@@ -2223,11 +2332,16 @@ from_ISO_8859_13_infos[129] = {
           o2(0xC5,0xBC),      o2(0xC5,0xBE),
      o3(0xE2,0x80,0x99),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_13 = {
+static const BYTE_LOOKUP
+from_ISO_8859_13 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_13_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_13 = {
+    "ISO-8859-13", "UTF-8", &from_ISO_8859_13, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_13_C2_offsets[64] = {
@@ -2362,11 +2476,16 @@ to_ISO_8859_13_infos[7] = {
      &to_ISO_8859_13_C5, &to_ISO_8859_13_E2,
                   UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_13 = {
+static const BYTE_LOOKUP
+to_ISO_8859_13 = {
     to_ISO_8859_10_offsets,
     to_ISO_8859_13_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_13 = {
+    "UTF-8", "ISO-8859-13", &to_ISO_8859_13, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_14_infos[129] = {
@@ -2436,11 +2555,16 @@ from_ISO_8859_14_infos[129] = {
           o2(0xC3,0xBD),      o2(0xC5,0xB7),
           o2(0xC3,0xBF),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_14 = {
+static const BYTE_LOOKUP
+from_ISO_8859_14 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_14_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_14 = {
+    "ISO-8859-14", "UTF-8", &from_ISO_8859_14, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_14_C2_offsets[64] = {
@@ -2652,11 +2776,16 @@ to_ISO_8859_14_infos[7] = {
      &to_ISO_8859_14_C5, &to_ISO_8859_14_E1,
                   UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_14 = {
+static const BYTE_LOOKUP
+to_ISO_8859_14 = {
     to_ISO_8859_14_offsets,
     to_ISO_8859_14_infos
 };
+static rb_transcoder
+rb_to_ISO_8859_14 = {
+    "UTF-8", "ISO-8859-14", &to_ISO_8859_14, 1, 1,
+    NULL, NULL,
+};
 
 static const struct byte_lookup* const
 from_ISO_8859_15_infos[129] = {
@@ -2726,11 +2855,16 @@ from_ISO_8859_15_infos[129] = {
           o2(0xC3,0xBD),      o2(0xC3,0xBE),
           o2(0xC3,0xBF),
 };
-const BYTE_LOOKUP
-rb_from_ISO_8859_15 = {
+static const BYTE_LOOKUP
+from_ISO_8859_15 = {
     from_ISO_8859_1_offsets,
     from_ISO_8859_15_infos
 };
+static rb_transcoder
+rb_from_ISO_8859_15 = {
+    "ISO-8859-15", "UTF-8", &from_ISO_8859_15, 3, 0,
+    NULL, NULL,
+};
 
 static const unsigned char
 to_ISO_8859_15_C2_offsets[64] = {
@@ -2840,10 +2974,47 @@ to_ISO_8859_15_infos[6] = {
       &to_ISO_8859_1_C3, &to_ISO_8859_15_C5,
      &to_ISO_8859_15_E2,              UNDEF,
 };
-const BYTE_LOOKUP
-rb_to_ISO_8859_15 = {
+static const BYTE_LOOKUP
+to_ISO_8859_15 = {
     to_ISO_8859_15_offsets,
     to_ISO_8859_15_infos
 };
-
+static rb_transcoder
+rb_to_ISO_8859_15 = {
+    "UTF-8", "ISO-8859-15", &to_ISO_8859_15, 1, 1,
+    NULL, NULL,
+};
+
+void
+Init_one_byte(void)
+{
+    rb_register_transcoder(&rb_from_ISO_8859_1);
+    rb_register_transcoder(&rb_from_ISO_8859_2);
+    rb_register_transcoder(&rb_from_ISO_8859_3);
+    rb_register_transcoder(&rb_from_ISO_8859_4);
+    rb_register_transcoder(&rb_from_ISO_8859_5);
+    rb_register_transcoder(&rb_from_ISO_8859_6);
+    rb_register_transcoder(&rb_from_ISO_8859_7);
+    rb_register_transcoder(&rb_from_ISO_8859_8);
+    rb_register_transcoder(&rb_from_ISO_8859_9);
+    rb_register_transcoder(&rb_from_ISO_8859_10);
+    rb_register_transcoder(&rb_from_ISO_8859_11);
+    rb_register_transcoder(&rb_from_ISO_8859_13);
+    rb_register_transcoder(&rb_from_ISO_8859_14);
+    rb_register_transcoder(&rb_from_ISO_8859_15);
+    rb_register_transcoder(&rb_to_ISO_8859_1);
+    rb_register_transcoder(&rb_to_ISO_8859_2);
+    rb_register_transcoder(&rb_to_ISO_8859_3);
+    rb_register_transcoder(&rb_to_ISO_8859_4);
+    rb_register_transcoder(&rb_to_ISO_8859_5);
+    rb_register_transcoder(&rb_to_ISO_8859_6);
+    rb_register_transcoder(&rb_to_ISO_8859_7);
+    rb_register_transcoder(&rb_to_ISO_8859_8);
+    rb_register_transcoder(&rb_to_ISO_8859_9);
+    rb_register_transcoder(&rb_to_ISO_8859_10);
+    rb_register_transcoder(&rb_to_ISO_8859_11);
+    rb_register_transcoder(&rb_to_ISO_8859_13);
+    rb_register_transcoder(&rb_to_ISO_8859_14);
+    rb_register_transcoder(&rb_to_ISO_8859_15);
+}
 /* Footprint (bytes): gross: 26788, saved: 3728, net: 23060 */
diff --git a/transcode.c b/transcode.c
index c915cc299d..e8c6545163 100644
--- a/transcode.c
+++ b/transcode.c
@@ -12,9 +12,9 @@
 
 #include "ruby/ruby.h"
 #include "ruby/encoding.h"
-
+#define PType (int)
 #include "transcode_data.h"
-
+#include <ctype.h>
 
 VALUE rb_str_tmp_new(long);
 VALUE rb_str_shared_replace(VALUE, VALUE);
@@ -23,168 +23,122 @@ VALUE rb_str_shared_replace(VALUE, VALUE);
  *  Dispatch data and logic
  */
 
-/* extern declarations, should use some include file here */
-extern const BYTE_LOOKUP rb_from_ISO_8859_1;
-extern const BYTE_LOOKUP rb_from_ISO_8859_2;
-extern const BYTE_LOOKUP rb_from_ISO_8859_3;
-extern const BYTE_LOOKUP rb_from_ISO_8859_4;
-extern const BYTE_LOOKUP rb_from_ISO_8859_5;
-extern const BYTE_LOOKUP rb_from_ISO_8859_6;
-extern const BYTE_LOOKUP rb_from_ISO_8859_7;
-extern const BYTE_LOOKUP rb_from_ISO_8859_8;
-extern const BYTE_LOOKUP rb_from_ISO_8859_9;
-extern const BYTE_LOOKUP rb_from_ISO_8859_10;
-extern const BYTE_LOOKUP rb_from_ISO_8859_11;
-extern const BYTE_LOOKUP rb_from_ISO_8859_13;
-extern const BYTE_LOOKUP rb_from_ISO_8859_14;
-extern const BYTE_LOOKUP rb_from_ISO_8859_15;
-
-extern const BYTE_LOOKUP rb_to_ISO_8859_1;
-extern const BYTE_LOOKUP rb_to_ISO_8859_2;
-extern const BYTE_LOOKUP rb_to_ISO_8859_3;
-extern const BYTE_LOOKUP rb_to_ISO_8859_4;
-extern const BYTE_LOOKUP rb_to_ISO_8859_5;
-extern const BYTE_LOOKUP rb_to_ISO_8859_6;
-extern const BYTE_LOOKUP rb_to_ISO_8859_7;
-extern const BYTE_LOOKUP rb_to_ISO_8859_8;
-extern const BYTE_LOOKUP rb_to_ISO_8859_9;
-extern const BYTE_LOOKUP rb_to_ISO_8859_10;
-extern const BYTE_LOOKUP rb_to_ISO_8859_11;
-extern const BYTE_LOOKUP rb_to_ISO_8859_13;
-extern const BYTE_LOOKUP rb_to_ISO_8859_14;
-extern const BYTE_LOOKUP rb_to_ISO_8859_15;
-
-extern const BYTE_LOOKUP rb_from_SHIFT_JIS;
-extern const BYTE_LOOKUP rb_from_EUC_JP;
-
-extern const BYTE_LOOKUP rb_to_SHIFT_JIS;
-extern const BYTE_LOOKUP rb_to_EUC_JP;
-
-extern void from_iso_2022_jp_transcoder_preprocessor(char**, char**, char*, char*,
-				struct transcoder_st *transcoder, struct transcoding*);
-extern void to_iso_2022_jp_transcoder_postprocessor(char**, char**, char*, char*,
-				struct transcoder_st *transcoder, struct transcoding*);
-
-/* declarations probably need to go into separate header file, e.g. transcode.h */
-
-/* todo: dynamic structure, one per conversion (stream) */
-
-/* in the future, add some mechanism for dynamically adding stuff here */
-#define MAX_TRANSCODERS 35  /* todo: fix: this number has to be adjusted by hand */
-static transcoder transcoder_table[MAX_TRANSCODERS];
-/* variable to work across register_transcoder and register_functional_transcoder */
-static int next_transcoder_position = 0;
-
-/* not sure why it's not possible to do relocatable initializations */
-/* maybe the code here can be removed (changed to simple initialization) */
-/* if we move this to another file???? */
-static void
-register_transcoder(const char *from_e, const char *to_e,
-    const BYTE_LOOKUP *tree_start, int max_output, int from_utf8)
+static st_table *transcoder_table, *transcoder_lib_table;
+
+#define TRANSCODER_INTERNAL_SEPARATOR '\t'
+
+static char *
+transcoder_key(const char *from_e, const char *to_e)
 {
-    if (next_transcoder_position >= MAX_TRANSCODERS) {
-	/* we are initializing, is it okay to use rb_raise here? */
-	rb_raise(rb_eRuntimeError /*change exception*/, "not enough transcoder slots");
-    }
-    transcoder_table[next_transcoder_position].from_encoding = from_e;
-    transcoder_table[next_transcoder_position].to_encoding = to_e;
-    transcoder_table[next_transcoder_position].conv_tree_start = tree_start;
-    transcoder_table[next_transcoder_position].max_output = max_output;
-    transcoder_table[next_transcoder_position].from_utf8 = from_utf8;
+    int to_len = strlen(to_e);
+    int from_len = strlen(from_e);
+    char *const key = xmalloc(to_len + from_len + 2);
+
+    memcpy(key, to_e, to_len);
+    memcpy(key + to_len + 1, from_e, from_len + 1);
+    key[to_len] = TRANSCODER_INTERNAL_SEPARATOR;
+    return key;
+}
 
-    next_transcoder_position++;
+void
+rb_register_transcoder(const rb_transcoder *tr)
+{
+    st_data_t k, val = 0;
+    const char *const from_e = tr->from_encoding;
+    const char *const to_e = tr->to_encoding;
+    char *const key = transcoder_key(from_e, to_e);
+
+    if (st_lookup(transcoder_table, (st_data_t)key, &val)) {
+	xfree(key);
+	rb_raise(rb_eArgError, "transcoder from %s to %s has been already registered",
+		 from_e, to_e);
+    }
+    k = (st_data_t)key;
+    if (st_delete(transcoder_lib_table, &k, &val)) {
+	xfree((char *)k);
+    }
+    st_insert(transcoder_table, (st_data_t)key, (st_data_t)tr);
 }
 
 static void
-register_functional_transcoder(const char *from_e, const char *to_e,
-    const BYTE_LOOKUP *tree_start, int max_output, int from_utf8,
-    void (*preprocessor)(char**, char**, char*, char*, transcoder*, transcoding*),
-    void (*postprocessor)(char**, char**, char*, char*, transcoder*, transcoding*))
+declare_transcoder(const char *to, const char *from, const char *lib)
 {
-    if (next_transcoder_position >= MAX_TRANSCODERS) {
-	/* we are initializing, is it okay to use rb_raise here? */
-	rb_raise(rb_eRuntimeError /*change exception*/, "not enough transcoder slots");
+    const char *const key = transcoder_key(to, from);
+    st_data_t k = (st_data_t)key, val;
+
+    if (st_delete(transcoder_lib_table, &k, &val)) {
+	xfree((char *)k);
     }
-    transcoder_table[next_transcoder_position].from_encoding = from_e;
-    transcoder_table[next_transcoder_position].to_encoding = to_e;
-    transcoder_table[next_transcoder_position].conv_tree_start = tree_start;
-    transcoder_table[next_transcoder_position].max_output = max_output;
-    transcoder_table[next_transcoder_position].from_utf8 = from_utf8;
-    transcoder_table[next_transcoder_position].conv_tree_start = tree_start;
-    transcoder_table[next_transcoder_position].preprocessor = preprocessor;
-    transcoder_table[next_transcoder_position].postprocessor = postprocessor;
-
-    next_transcoder_position++;
+    st_insert(transcoder_lib_table, (st_data_t)key, (st_data_t)lib);
 }
 
-static void
-init_transcoder_table(void)
+#define MAX_TRANSCODER_LIBNAME_LEN 64
+static const char transcoder_lib_prefix[] = "enc/trans/";
+
+void
+rb_declare_transcoder(const char *enc1, const char *enc2, const char *lib)
 {
-    register_transcoder("ISO-8859-1",  "UTF-8", &rb_from_ISO_8859_1, 2, 0);
-    register_transcoder("ISO-8859-2",  "UTF-8", &rb_from_ISO_8859_2, 2, 0);
-    register_transcoder("ISO-8859-3",  "UTF-8", &rb_from_ISO_8859_3, 2, 0);
-    register_transcoder("ISO-8859-4",  "UTF-8", &rb_from_ISO_8859_4, 2, 0);
-    register_transcoder("ISO-8859-5",  "UTF-8", &rb_from_ISO_8859_5, 3, 0);
-    register_transcoder("ISO-8859-6",  "UTF-8", &rb_from_ISO_8859_6, 2, 0);
-    register_transcoder("ISO-8859-7",  "UTF-8", &rb_from_ISO_8859_7, 3, 0);
-    register_transcoder("ISO-8859-8",  "UTF-8", &rb_from_ISO_8859_8, 3, 0);
-    register_transcoder("ISO-8859-9",  "UTF-8", &rb_from_ISO_8859_9, 2, 0);
-    register_transcoder("ISO-8859-10", "UTF-8", &rb_from_ISO_8859_10, 3, 0);
-    register_transcoder("ISO-8859-11", "UTF-8", &rb_from_ISO_8859_11, 3, 0);
-    register_transcoder("ISO-8859-13", "UTF-8", &rb_from_ISO_8859_13, 3, 0);
-    register_transcoder("ISO-8859-14", "UTF-8", &rb_from_ISO_8859_14, 3, 0);
-    register_transcoder("ISO-8859-15", "UTF-8", &rb_from_ISO_8859_15, 3, 0);
-    register_transcoder("UTF-8", "ISO-8859-1",  &rb_to_ISO_8859_1, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-2",  &rb_to_ISO_8859_2, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-3",  &rb_to_ISO_8859_3, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-4",  &rb_to_ISO_8859_4, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-5",  &rb_to_ISO_8859_5, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-6",  &rb_to_ISO_8859_6, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-7",  &rb_to_ISO_8859_7, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-8",  &rb_to_ISO_8859_8, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-9",  &rb_to_ISO_8859_9, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-10", &rb_to_ISO_8859_10, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-11", &rb_to_ISO_8859_11, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-13", &rb_to_ISO_8859_13, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-14", &rb_to_ISO_8859_14, 1, 1);
-    register_transcoder("UTF-8", "ISO-8859-15", &rb_to_ISO_8859_15, 1, 1);
-
-    register_transcoder("SHIFT_JIS", "UTF-8",   &rb_from_SHIFT_JIS, 3, 0);
-    register_transcoder("EUC-JP", "UTF-8",      &rb_from_EUC_JP, 3, 0);
-    register_transcoder("UTF-8", "SHIFT_JIS",   &rb_to_SHIFT_JIS, 2, 1);
-    register_transcoder("UTF-8", "EUC-JP",      &rb_to_EUC_JP, 2, 1);
-    register_functional_transcoder("ISO-2022-JP", "UTF-8", &rb_from_EUC_JP,
-				   8, 0, &from_iso_2022_jp_transcoder_preprocessor, NULL);
-    register_functional_transcoder("UTF-8", "ISO-2022-JP", &rb_to_EUC_JP,
-				   8, 1, NULL, &to_iso_2022_jp_transcoder_postprocessor);
-
-    register_transcoder(NULL, NULL, NULL, 0, 0);
+    if (!lib || strlen(lib) > MAX_TRANSCODER_LIBNAME_LEN) {
+	rb_raise(rb_eArgError, "invalid library name - %s",
+		 lib ? lib : "(null)");
+    }
+    declare_transcoder(enc1, enc2, lib);
+    declare_transcoder(enc2, enc1, lib);
 }
 
-static int
-encoding_equal(const char* encoding1, const char* encoding2)
+static void
+init_transcoder_table(void)
 {
-    return 0==strcasecmp(encoding1, encoding2);
+    rb_declare_transcoder("ISO-8859-1",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-2",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-3",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-4",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-5",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-6",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-7",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-8",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-9",  "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-10", "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-11", "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-13", "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-14", "UTF-8", "single_byte");
+    rb_declare_transcoder("ISO-8859-15", "UTF-8", "single_byte");
+    rb_declare_transcoder("SHIFT_JIS",   "UTF-8", "japanese");
+    rb_declare_transcoder("EUC-JP",      "UTF-8", "japanese");
+    rb_declare_transcoder("ISO-2022-JP", "UTF-8", "japanese");
 }
 
-static transcoder*
+#define encoding_equal(enc1, enc2) (strcasecmp(enc1, enc2) == 0)
+
+static rb_transcoder *
 transcode_dispatch(const char* from_encoding, const char* to_encoding)
 {
-    transcoder *candidate = transcoder_table;
-    
-    for (candidate = transcoder_table; candidate->from_encoding; candidate++) {
-	if (encoding_equal(from_encoding, candidate->from_encoding)
-	    && encoding_equal(to_encoding, candidate->to_encoding)) {
-		return candidate;
+    char *const key = transcoder_key(from_encoding, to_encoding);
+    st_data_t k, val = 0;
+
+    k = (st_data_t)key;
+    if (!st_lookup(transcoder_table, k, &val) &&
+	st_delete(transcoder_lib_table, &k, &val)) {
+	const char *const lib = (const char *)val;
+	int len = strlen(lib);
+	char path[sizeof(transcoder_lib_prefix) + MAX_TRANSCODER_LIBNAME_LEN];
+
+	xfree((char *)k);
+	if (len > MAX_TRANSCODER_LIBNAME_LEN) return NULL;
+	memcpy(path, transcoder_lib_prefix, sizeof(transcoder_lib_prefix) - 1);
+	memcpy(path + sizeof(transcoder_lib_prefix) - 1, lib, len + 1);
+	if (!rb_require(path)) return NULL;
+	if (!st_lookup(transcoder_table, (st_data_t)key, &val)) {
+	    /* multistep logic, via UTF-8 */
+	    if (!encoding_equal(from_encoding, "UTF-8") &&
+		!encoding_equal(to_encoding, "UTF-8") &&
+		transcode_dispatch("UTF-8", to_encoding)) {  /* check that we have a second step */
+		return transcode_dispatch(from_encoding, "UTF-8"); /* return first step */
+	    }
+	    return NULL;
 	}
     }
-    /* multistep logic, via UTF-8 */
-    if (!encoding_equal(from_encoding, "UTF-8")
-	&& !encoding_equal(to_encoding, "UTF-8")
-	&& transcode_dispatch("UTF-8", to_encoding)) {  /* check that we have a second step */
-	    return transcode_dispatch(from_encoding, "UTF-8"); /* return first step */
-    }
-    return NULL;
+    return (rb_transcoder *)val;
 }
 
 
@@ -194,8 +148,8 @@ transcode_dispatch(const char* from_encoding, const char* to_encoding)
 static void
 transcode_loop(char **in_pos, char **out_pos,
 	       char *in_stop, char *out_stop,
-	       transcoder *my_transcoder,
-	       transcoding *my_transcoding)
+	       const rb_transcoder *my_transcoder,
+	       rb_transcoding *my_transcoding)
 {
     char *in_p = *in_pos, *out_p = *out_pos;
     const BYTE_LOOKUP *conv_tree_start = my_transcoder->conv_tree_start;
@@ -280,7 +234,7 @@ transcode_loop(char **in_pos, char **out_pos,
  */
 
 static char *
-str_transcoding_resize(transcoding *my_transcoding, int len, int new_len)
+str_transcoding_resize(rb_transcoding *my_transcoding, int len, int new_len)
 {
     VALUE dest_string = my_transcoding->ruby_string_dest;
     rb_str_resize(dest_string, new_len);
@@ -298,8 +252,8 @@ str_transcode(int argc, VALUE *argv, VALUE *self)
     const char *from_e, *to_e;
     int from_encidx, to_encidx;
     VALUE from_encval, to_encval;
-    transcoder *my_transcoder;
-    transcoding my_transcoding;
+    rb_transcoder *my_transcoder;
+    rb_transcoding my_transcoding;
     int final_encoding = 0;
 
     if (argc<1 || argc>2) {
@@ -307,6 +261,7 @@ str_transcode(int argc, VALUE *argv, VALUE *self)
     }
     if ((to_encidx = rb_to_encoding_index(to_encval = argv[0])) < 0) {
 	to_enc = 0;
+	to_encidx = 0;
 	to_e = StringValueCStr(to_encval);
     }
     else {
@@ -405,7 +360,7 @@ str_transcode(int argc, VALUE *argv, VALUE *self)
     }
     /* set encoding */
     if (!to_enc) {
-	to_encidx = rb_enc_replicate(to_e, rb_ascii8bit_encoding());
+	to_encidx = rb_define_dummy_encoding(to_e);
     }
     *self = dest;
 
@@ -467,7 +422,10 @@ rb_str_transcode(int argc, VALUE *argv, VALUE str)
 void
 Init_transcode(void)
 {
+    transcoder_table = st_init_strcasetable();
+    transcoder_lib_table = st_init_strcasetable();
     init_transcoder_table();
+
     rb_define_method(rb_cString, "encode", rb_str_transcode, -1);
     rb_define_method(rb_cString, "encode!", rb_str_transcode_bang, -1);
 }
diff --git a/transcode_data.h b/transcode_data.h
index 3131877a41..add954ef74 100644
--- a/transcode_data.h
+++ b/transcode_data.h
@@ -22,12 +22,9 @@ typedef struct byte_lookup {
     const struct byte_lookup *const *info;
 } BYTE_LOOKUP;
 
-#ifdef TRANSCODE_DATA
+#ifndef PType
 /* data file needs to treat this as a pointer, to remove warnings */
 #define PType (const BYTE_LOOKUP *)
-#else
-/* in code, this is treated as just an integer */
-#define PType (int)
 #endif
 
 #define NOMAP   (PType 0x01)   /* single byte direct map */
@@ -56,23 +53,26 @@ typedef struct byte_lookup {
 
 /* dynamic structure, one per conversion (similar to iconv_t) */
 /* may carry conversion state (e.g. for iso-2022-jp) */
-typedef struct transcoding {
+typedef struct rb_transcoding {
     VALUE ruby_string_dest; /* the String used as the conversion destination,
 			       or NULL if something else is being converted */
-    char *(*flush_func)(struct transcoding*, int, int);
-} transcoding;
+    char *(*flush_func)(struct rb_transcoding*, int, int);
+} rb_transcoding;
 
 /* static structure, one per supported encoding pair */
-typedef struct transcoder_st{
+typedef struct rb_transcoder {
     const char *from_encoding;
     const char *to_encoding;
     const BYTE_LOOKUP *conv_tree_start;
     int max_output;
     int from_utf8;
     void (*preprocessor)(char**, char**, char*, char*,
-				struct transcoder_st *transcoder, struct transcoding*);
+			 struct rb_transcoder *, struct rb_transcoding *);
     void (*postprocessor)(char**, char**, char*, char*,
-				 struct transcoder_st *transcoder, struct transcoding*);
-} transcoder;
+			 struct rb_transcoder *, struct rb_transcoding *);
+} rb_transcoder;
+
+void rb_declare_transcoder(const char *enc1, const char *enc2, const char *lib);
+void rb_register_transcoder(const rb_transcoder *);
 
 #endif /* RUBY_TRANSCODE_DATA_H */